For a number of years, it has been known that various microorganisms exhibit biological activity so as to be useful to control plant diseases. Although progress has been made in the field of identifying and developing biological pesticides for controlling various plant diseases of agronomic and horticultural importance, most of the pesticides in use are still synthetic compounds. Many of these chemical fungicides are classified as carcinogens by the EPA, are toxic to wildlife and other non-target species. In addition, pathogens may develop resistance to chemical pesticides (see, e.g., Schwinn et al., p. 244, ADVANCES IN PLANT PATHOLOGY: PHYTOPHTHORA INFESTANS, THE CAUSE OF LATE BLIGHT OF POTATO (Academic Press, San Diego 1991).
Every year 250-300 million dollars of chemical pesticides are used to control corn rootworm infestations. Many of these chemical pesticides are toxic to humans, wildlife and other nontarget species. Also some have been found in the ground water. New chemical insecticides cost $100 million to develop.
Biological control offers an attractive alternative to synthetic chemical fungicides. Biopesticides (living organisms and the naturally produced compounds produced by these organisms) can be safer, more biodegradable, and less expensive to develop.
Screening programs have identified certain Bacillus spp. (Bacillus spp. includes B. subtilis, B. cereus, B. mycoides, B. thuringiensis) strains that exhibit antifungal activity. (See e.g Stabb et al. (1990) Applied Environ. Microbiol. 60: 4404-4412). These strains have been shown to produce zwittermicin-A and or kanosamine (Milner et al. (1996) Appl. Environ. Microb. 62: 3061-3066), two antibiotic agents that are effective against the soil borne disease damping off, caused by Phytophthora medicaginis, P. nicotianae, P. aphanidermatum or Sclerotinia minor (See Stabb et al., supra). Zwittermicin-A is a water soluble, acid stable linear aminopolyol molecule (see, He et al, (1994) Tetra. Lett. 35 (16) 2499-2502.
U.S. Pat. No. 5,049,379 to Handelsman et al. describes how zwittermicin-A produces damping off in alfalfa and soybeans. When the seed was coated with B. cereus ATCC 53522, the pathogenic activity of root rot fungus is inhibited. Similarly application of spore-based formulations of certain B. cereus strains to soybean seeds or the soil surrounding the seeds has been shown to improve soybean yield at field sites. (See, Osburne et al (1995) Am. Phytopathol. Soc. 79(6): 551-556). Methods of applying biopesticides are well known in the art and include, for example, wettable powders, dry flowables, microencapsulation of effective agents, liquid or solid formulations of antibiotic fractions from suitable cultures. (See e.g., U.S. Pat. No. 5,061,495 to Rossall or U.S. Pat. No. 5,049,379 to Handelsman).
Smith et al. (1993) Plant Disease 77(2) 139-142 report that the activity of the soil-borne fungus, Pythium aphanidermatum, that causes cottony cucumber leak can be suppressed using zwittermicin-producing B. cereus strain UW85. Leifert et al. (1995) J. Appl. Bacteriol. 78: 97-108 report that the production of anti-Botrytis and anti-Alternaria antibiotics by two Bacillus strains, B. subtilis CL27 and B. pumilis CL 45. The whole broth and cell-free filtrates were active against Botrytis and Alternaria in in vitro tests and were active against Botrytis in in vivo small plant tests on Astilbe. Leifert et al. (1997) U.S. Pat. No. 5,597,565 disclose B. subtilis, B. pumilis, and B. polymyxa that are particularly effective at inhibiting post harvest disease causing fungi. They also disclose the presence of antibiotics produced in the cell-free culture filtrate and their activity at different pH values, but they do not identify these compounds.
Rossall (1994) U.S. Pat. No. 5,344,647 discloses Bacillus subtilis strains with broad anti-fungal activity. Sholberg et al. (1995) Can. J. Microbiol. 41: 247-252, Swinburne et al. (1975) Trans. Brit. Mycol. Soc. 65: 211-217, Singh and Deverall (1984) Trans. Br. Mycol. Soc. 83: 487-490, and Ferreira, et al. (1991) Phytopathology 81: 283-287. Baker et al. (1983) Phytopathology 73: 1148-1152 disclose the use of Bacillus spp. and Bacillus subtilis as biocontrol agents of fungal plant pathogens. Baker et al. (1983) Phytopathology 73: 1148-1152 also report on an antifungal Bacillus subtilis for use on plant pathogens. Pusey et al. (1988) Plant Dis. 72: 622-626, Pusey and Robins (U.S. Pat. No. 5,047,239), and McKeen et al. (1986) Phytopathology 76: 136-139 disclose control of post harvest fruit rot using B. subtilis. McKeen et al, supra, have shown that antibiotics similar to the low molecular weight iturin cyclic polypeptides contribute to this fungicidal activity of B. subtilis.
Liu et al. (1995) U.S. Pat. No. 5,403,583 disclose a Bacillus megaterium, ATCC 55000 and a method to control the fungal plant pathogen, Rhizoctonia solani. Islam and Nandi (1985) Journal of Plant Diseases and Protection 92(3): 241-246 disclose a Bacillus megaterium with antagonism to Drechslera oryzae, the causal agent of rice brown spot. The same authors, Islam and Nandi (1985) Journal of Plant Diseases and Protection 92(3) 233-240 also disclose in-vitro antagonism of B. megaterium against Drechslera oryzae, Alternaria alternata and Fusarium roseum. They discuss three components in the culture filtrate. The most active antibiotic was highly soluble in water and methanol with a UV peak at 255 nm and a shoulder at 260 nm, which proved to be a polyoxin-like lipopeptide. Cook ((1987) Proceedings Beltwide Cotton Production--Mechanization Research Conference, Cotton Council, Memphis, p. 43-45) discloses the use of a suspension of Bacillus megaterium to reduce the number of cotton plants killed by Phymatotrichum omnivorum, a cause of cotton root rot.
Antibiotic production of B. megaterium has been recorded by Berdy (CRC Handbook of Antibiotic Compounds, Vols. I-XIV, (CRC Press, Inc. Boca Raton, Fla. 1980-87) who reports production of low-mammalian toxic peptide antibiotics such as ansamitocin-PDM-O, bacimethrin, megacin, pentapeptide, homopeptides.
Bacilli are known to produce antifungal and antibacterial secondary metabolites (Korzybski et al. (1978)). University of Wisconsin and Cornell researchers have identified a novel fungicidal compound, zwittermicin A, produced by Bacillus sp. (He et al. (1994) Tetra. Lett. 35(16):2499-2502). A second fungicidal metabolite produced by the same strain was recently identified as the known amino-sugar, kanosamine (Milner et al. (1996) Appl. Environ. Microb. 62:3061-3065).
Another group of previously described Bacillus metabolites are the cyclic lipopeptides of the iturin class, some of which are potent fingicidal agents. These agents consist of a cyclic octapeptide with seven .alpha.-amino acids and one .beta.-amino acid with an aliphatic side chain. There are several groups of iturins that differ in order and content of the amino acid sequence. These are shown in Table 1 below. Generally, a suite of related molecules is produced with differences in the length and branching of the aliphatic amino acid residue. When tested against Saccharomyces cerevesiae, mycosubtilin was found to be the most active agent (LC50=10 .mu.g/mL) followed by iturin-A and bacillomycin L (both haviing an LC50=30 .mu.g/mL) (Beeson et al. (1979) J. Antibiotics 32(8):828-833). The mode of action of these cyclic lipopeptides has been reported to be due to interaction with fungal membranes creating transmembrane channels that permit release of vital ions (Latoud et al. (1986) Biochem. Biophys. Acta 856:526-535). Iturin-C is inactive against fungi including Penicillium chrysogenum (Peypoux et al. (1978) Tetrahedron 34:1147-1152).
TABLE 1 Structures of the iturin family of antibiotics Antibiotic L-Asz(X1) X4 X5 X6 X7 Iturin A L-Asn L-Gln L-Pro D-Asn L-Ser Iturin C L-Asp L-Gln L-Pro D-Asn L-Ser Bacillo- L-Asn L-Pro L-Glu D-Ser L-Thr mycin D Bacillo- L-Asp L-Ser L-Gln D-Ser L-Thr mycin L Bacillo- L-Asn L-Gln L-Pro D-Asn L-Thr mycin F Myco- L-Asn L-Gln L-Pro D-Ser L-Asn subtilin
##STR1##
A research group at the USDA has investigated the structure/activity relationship of the iturins by synthesizing a number of analogs differing in the amino acid chain length. The researchers reported that the activity of the iturins increased with the length of the fatty acid side chain and the terminal branching in the order iso&gt;normal&gt;anteiso (Bland et al. (1995) Proc. Plant Growth Regulation Soc. Am. 22nd: 105-107). They also state that the "amounts of iturins obtained from natural production are inadequate to be commercially viable" based on their work with a number of iturin producing strains of Bacillus.
Another groups of cyclic lipopeptides isolated from B. cereus are the plipastatins. These compounds are a family of acylated decapeptides, the structures of which are shown in FIG. 1 (Nishikiori et al. (1986) J. Antibiotics 39(6):755-761). These compounds were originally isolated as inhibitors of porcine pancreatic phospholipase A.sub.2 (Umezawa et al. (1986) J. Antibiotics 39(6):737-744), but were later found to inhibit some plant pathogenic fungi including Botrytis, Pyricularia and Alternaria (Yamada et al. (1990) Nippon Noyaku Gakkaishi 15(1):95-96). Yamada also reported a synergistic effect observed between iturin A and the plipastatins, both produced by the same B. subtilis strain.
Work has been carried out on fermentation improvements to increase production of the iturins in both liquid (Phae and Shoda (1991) J. Ferment. Bioeng. 71:118-121); Ohno et al. (1993) J. Ferment. Bioeng. 75:463-465) and solid state fermentations (Ohno et al. (1992) Biotech. Lett. 14(9):817-822; Ohno et al. (1995) J. Ferment. Bioeng. 5:517-519). There is a report of synergy between the closely related surfactins, that are themselves inactive, and the iturins produced by the same B. subtilis strain (Hiraoka et al. (1992) J. Gen. Appl. Microbiol. 38:635-640). The nucleotide sequence for the gene that co-regulates biosynthesis of iturin A and surfactin has been published (Huang et al. (1993) J. Ferment. Bioeng. 76(6):445-450). Field work on iturin-producing strains has concentrated on soil treatment for control of Rhizoctonia (Asaka and Shoda (1996) Appl. Environ. Microbiol. 62:4081-4085) and foliar field applications of iturins have not been reported.
Another cyclic lipopeptide compound produced by B. subtilis is surfactin, which possesses an exceptional surfactant activity (Kaninuma et al. (1969) Agric. Biol. Chem. 33:973-976). Surfactin contains a C14 or C15 .beta.-hydroxy fatty acid linked by a lactone ring to a heptapeptide moiety with a LLDLLDL (SEQ ID NO:1) sequence (Arima et al. (1968) Biochem. Biophys. Res. Commun. 31:488-494. Sandrin et al. ((1990) Biotechnol. Appl. Biochem. 12:370-375) found B. subtilis strains that produced both surfactin and iturin A, the bacillomycins F and L and mycosubtilin.
The novel microorganism AQ713 discovered by the inventors, previously thought to be a strain of Bacillus megaterium and now identified as a strain of Bacillus subtilis, produces A iturins, plipastatins and surfactins. Production of this combination of lipopeptides by a microorganism has not been previously reported. In addition, the inventors have discovered that AQ713 also produces a newly described group of compounds designated as "agrastatins." The combination of all three of the above known compounds with the novel agrastatins is also novel.
One commonly used biopesticide is the gram positive bacterium Bacillus thuringiensis. Pesticidal B. thuringiensis strains are known to produce crystal proteins during sporulation, which are specifically toxic to certain orders and species of insects and nematodes (See, e.g., U.S. Pat. Nos. 4,999,192 and 5,208,017). Proteinaceous endotoxins produced by B. thuringiensis also act as insecticidal agents against corn rootworm and other beetles (e.g. U.S. Pat. No. 5,187,091; Johnson, T. J. et al. (1993), J. Economic Entomology 86:330-333). B. thuringiensis endotoxins have been shown to be effective as purified crystals, washed cell pellets, and expressed proteins. Warren et al. (WO 96/10083), discloses non-endotoxin proteins produced during the vegetative stage of Bacillus cereus and B. thuringiensis. These vegetative proteins, called Vip1 and Vip2 have potent activity against corn rootworm (northern and western) (Estruch et al. (1997), Nature Biotechnology 15:137-141 and Mullins et al. (1997), Appl Environ. Microbiol. 63, (in press).
One B. thuringiensis thermostable metabolites, termed beta-exotoxin has also been shown to have pesticidal properties. Burgjeron and Biache (1979), Entomophaga 11:279-284 report a beta exotoxin that is active against Colorado potato beetle (Leptinotarsa decemlineata). In addition, the known B. thuringiensis beta-exotoxins exhibits non-specific pesticidal activity, killing not only nematodes, but also flies, armyworms, mites, and corn rootworms. Sigma exotoxin has a structure similar to beta-exotoxin, and is active against Colorado potato beetle (Argauer et al. (1991) J. Entomol. Sci. 26:206-213). Alpha-exotoxin is toxic against larvae of Musca domestica (Cluthy (1980) FEMS Microbiol. Lett. 8:1-7). Gamma-exotoxins are various proteolytic enzymes, chitinases and proteases. The toxic effects of gamma exotoxins are only expressed in combination with beta-exotoxin or delta-endotoxin. Forsberg et al. (1976) "Bacillus thuringiensis: Its effects in Environmental Quality," National Research Council of Canada. Stonard et al. (1994) ACS Symposium Series 551: 25 report a water-soluble secondary metabolite active against corn rootworm in the supernatant of a Bacillus cereus strain.
There are no documented strains of Bacillus with bacteriacidal, fungicidal and corn rootworm activity. There are no known metabolites produced by Bacillus subtilis that are of less than 10,000 molecular weight and are extractable in a non-polar solvent.